MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202641069777 A) filed by Indian Institute Of Technology Madras Iit Madras on June 03, 2026, for System And Method For Selecting A Robust Model For Estimating Thermal Errors In Machine Tool.
Inventors include Devadula, Sivasrinivasu; and Gunjan, Shashi Bhushan.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: Disclosed is a system (100) and a method (400) for selecting a robust model for estimating thermal errors in a machine tool (104). The method (400) includes receiving temperature data and corresponding Tool Center Point drift measurements from machine tool sensors over an operational period. Multiple candidate Machine Learning (109-1)/Deep Learning models (109-2) are trained using the received data. Degraded temperature data is generated using a sensor network degradation model integrated with a Fractional Sobol-Quasi Monte Carlo (FS-QMC) model to simulate sensor degradation scenarios. A simulated performance metric is computed and used to derive a limit state function for each candidate model relative to a threshold metric. Model reliability and output information are determined, and a robust candidate model is selected based on the reliability and the output information for estimating thermal errors, enabling a reliability-driven, explainable framework for robust selection of TEPM under sensor degradation. Representative Figure: FIG. 4
Disclaimer: Curated by HT Syndication.
